The EMBO Journal vol.9 no.1 1 pp.3447-3452, 1990

A feedback control element near the transcription startsite of the maize Shrunken gene determines promoteractivity

Christoph Maas, Stefan Schaal andWolfgang Werr

Institut fir Genetik der Universitat zu Koin Weyertal 121, 5000 Koin41, FRG

Communicated by P.Starlinger

The transcriptional activity of the Shrunken (Sh)promoter of Zea mays was monitored in transientexpression assays using the neomycin phosphotransferase(NPT) H gene as a reporter in maize suspension proto-plasts. Shortly after transfection, expression of thischimeric NPIIM gene was negatively affected by highextracellular sucrose concentrations in the protoplastcultivation medium. However, 3-5 days after trans-fection an up to 405-fold increase in NPMI activity wasobserved. This could be blocked by dichlorobenzonitril(DCB) an inhibitor of cellulose biosynthesis. In theanalysis of promoter deletions 20 bp upstream of the Shtranscription start site were sufficient to reproduce theexpression profile and the activity of the full promoter.Surprisingly this start sequence does not include thenatural TATA-box.Key words: feedback control/Shnken gene/transcription/Zeamays

IntroductionIn most plants, including Zea mays, carbon fixation in thephotosynthetically active plant organs is followed byconversion of the photoassimilates to sucrose which istransported to the sink tissues (reviews: Akazawa andOkamoto, 1980; Giaquinta, 1983; Stitt and Quick, 1989).Many steps in carbohydrate metabolism have been identifiedand analysed at the level of enzyme activity. However, littleis known about the transcriptional control of the encodinggenes and its potential effects on the partitioning of sucrosebetween the various sink organs.The unloading of sucrose from the phloem into the maize

endosperm and subsequent starch synthesis have beenextensively studied on the enzymatic and metabolite level(Tsai et al., 1970; Griffith et al., 1987; Doehlert et al.,1988). Several enzyme mutants are known which affecteither sugar concentrations or starch synthesis in theendosperm tissue (review: Creech, 1965) and some of theencoding genes have been cloned and analysed (Kl6sgenet al., 1986; McCarty et al., 1986). The enzyme sucrosesynthase, encoded by the Shrunken gene on chromosome9 ofZ.mays, catalyses the reaction sucrose + UDP - UDP-glucose + fructose. Homozygous mutants of this geneaccumulate only 40% of the starch found in wild-type plants(Chourey and Nelson, 1976). Therefore, either synthesis or

cleavage of sucrose is a rate limiting step prior to thedeposition of carbohydrates in starch. The Shrunken genehas been cloned (Geiser et al., 1982) and analysed (Werret al., 1985). In situ studies have shown that in theendosperm, expression is confined to cells activelysynthesizing starch (Heinlein and Starlinger, 1989; Chen andChourey, 1989). The Shrunken gene is also transcribed inroots and shoots of etiolated maize seedlings with thetranscript level increasing during anaerobiosis (Springeret al., 1986). Recent in situ experiments demonstrate celltype specific transcription, especially in the root tip, andincreased transcription in the vascular cylinder of anaero-bically treated maize roots (Rowland et al., 1989).The transient gene expression experiments described here

indicate that transcription of genes under control of theShrunken promoter in maize protoplasts is repressed by highlevels of extracellular sucrose and activated when theprotoplasts commence cell wall regeneration. Deletionanalysis of the upstream sequences has identified oneimportant regulatory region close to the transcription start.

ResultsTranscription from the Shrunken promoter in maizeprotoplasts is delayedThe activity of the Shrunken (Sh) promoter in maizesuspension protoplasts was tested using a chimeric constructcontaining a 2014 bp Sh upstream sequence and 42 bp ofthe untranslated leader sequence encoded by exon 1 clonedin front of the neomycin phosphotransferase marker gene(details of the chimeric SP2014+42 NPT gene are shownin Figure la). The expression of SP2014+42 NPT wasanalysed at various time points after transfection. The NPTIIactivity was low in the 48 h following transfection butincreased sharply afterwards (Figure lb). The total increasein NPTII activity observed between 1-2 and 5 days aftertransfection varied in the range 340- to 405-fold in sixindependent experiments. In cotransfection experimentsusing a chimeric chloramphenicol transacetylase (CAT) geneunder the control of the cauliflower mosaic virus (CaMV)35S promoter (pRT1O1CAT, Prols et al., 1988) the activityprofile of the Sh promoter in maize protoplasts was incontrast to that of the 35S promoter. Highest CAT activitywas observed between 19 and 24 h after transfection(Figure lb) demonstrating that transient expression of theintroduced chimeric genes was not delayed because ofprotoplast recovery. Similarly, NPTII expression wasmaximal between 19 and 24 h after transfection if the markergene was transcribed from either the CaMV 35S (pK5, seeFigure la and b; Kraus, 1988) or the nopaline synthase (nos)promoter (pLGV1 103, Hain et al., 1985; data not shown).Therefore the time-lag observed with the Sh promoter is alsonot explained by a peculiarity of the NPTII marker gene.

© Oxford University Press 3447

C.Maas, S.Schaal and W.Werr

b_

TAT*-box

_ FK~rs iLd SP Z14+42 ffPT+42

28 SP29+4Z NPT

-20-2 . . rrr ." SP28+6 FPT

NIvr SP 2014-47 core NPT

Htu

3Cm1Cm

- 47

NPTFI a~, 35S core NPT

NPr rit..c p IIll4-5

- SU

compIete AP3f1 t2CaftU 35s pRTiOl CRTpromiotergj T9(42CA.4moter7E pRTIBt4+42)CRT

+4 + 42

Cm

C14 (D 0 CDC4JCDII I I3

SP2014+42 NPT

hours

35S - CAT(pRT101 CAT)

0gO C,4u cs V r- 0) hours

_:I* __ 35S-NPT__ ~ ~(pK5)

Fig. 1. (a) Schematic representation of chimeric gene constructions transfected into maize suspension protoplasts. The numbering of the sucrosesynthase upstream fragments indicates the distance of the terminal nucleotides from the major transcription start site (+1) used in the maizeendosperm (Werr et al., 1985). The CaMV 35S core promoter fragment extends to the EcoRV restriction site at position -90. (b) Activity of the Shpromoter in comparison with the CaMV 35S promoter. Protoplasts of one large scale transfection were cultivated, isolated at different time pointsafter transfection and monitored for expression of the chimeric SP2014+42 NPT and cotransfected 35S-CAT (pRTlOlCAT, Prols et al., 1988).The NPTII activity profile observed with the chimeric 35S-NPT gene (pK5) in a different transfection experiment is shown in comparison below.NPTII: protein extract from a tobacco plant stably transformed with a chimeric NPTII gene, CAT: chloramphenicol transacetylase enzyme Cm:chloramphenicol, lCm and 3Cm: acetylated reaction products.

Suc GW-

\ </Fru/..

/' Fr

--1 1?0 2 3 4 5

pK5 SP2014+42 NPT

175 * FSUC

*117 j jGlu

*: 58 FrumM[

_ ~~~~~GlutSuc - Fru

175 mM

Suc Glu Fru1

Fru. / /

Suc? /Glu

..41~I

0 1 2 3 4 5DAT

a. 4iii..

k .,

175mM

2

3

4

5

J!Wa-

Fig. 2. (a) The Sh promoter is inhibited by high extracellular sucrose concentrations. The diagram indicates sucrose hydrolysis and accumulation ofglucose and fructose in the protoplast medium. The higher differences observed between 2 and 4 days are due to faster hydrolysis of sucrose indifferent experimental series. Beside the diagram the NPTII activity is shown observed 2 days after transfection of protoplasts with eitherSP2014+42 NPT or pK5 (35S promoter) and cultivated in media with decreasing sucrose concentrations. Expression of the CaMV 35S promoter(pK5) was constant, while the activity of the Sh promoter increased at lower sucrose concentrations. The inhibition of the Sh promoter was notobserved at 175 mM concentrations of glucose, fructose or glucose/fructose (each 175 mM). (b) Transient expression of SP2014+42 NPT inprotoplasts cultivated on different carbohydrate sources. One large scale transfection experiment was split into three aliquots which were plated on

media containing 175 mM sucrose, glucose or fructose and cultured for the time indicated beside the autoradiographs. Note that sucrose ishydrolysed during the cultivation period (see Figure 2a). The NPTII activity values in the diagram have been nornalized on the level determined insucrose-containing medium at day 1 after transfection. DAT: days after transfection.

The Sh promoter is inhibited by high extracellularsucrose levelsAnalysis of the carbohydrate composition in the mediumduring protoplast cultivation indicated that the sucrose

supplied as a carbohydrate source was completely hydrolysed

within the 5 day cultivation period. As depicted in Figure 2aglucose and fructose accumulated to equimolar concentra-tions. Therefore transient expression of SP2014+42 NPTwas compared after 2 days cultivation in mediumsupplemented with sucrose, glucose or fructose (each 175

3448

-aa.

-2Z814

161 -

80 -

( 4)E.0

o

v-)13 -

> 400-

C.)

mC 300-

z 200-z

100 -

20 -

44t

Aim.,

0

-28D4

Feedback control element of maize

Table I. Summary of transient expression experiments with variouspromoter NPTII constructions.

5 days2 days

Glu + FruSuc

SP20-42-DCB

1 2 3 4 5

NPT+DCB

1 2 3 4 5 DAT

TAT&-b.x

-2U62-21 SW>UNT

-2@_ /{WIP

.1142

340 - 405 16 - 25

240- 320 8 - 20

10- 26

__Q " ,m-4WEflPT 15- 30-2614 .-

7-12

6 - 7.5

u55MM T 1.1-1 .7 1.0 -1 .5

Transient expression of Sh promoter deletions was compared aftercultivation in glucose or sucrose containing medium and at day 2 or 5after transfection. The numbers are calculated from differentexperiments and give relative values nonnalized on the NPTII activityobserved after 2 days cultivation on 175 mM sucrose media (= 1).

mM, Figure 2a). Transcription from the Sh promoter as

estimated by comparison of NPTH activities was - 20-foldhigher in protoplasts cultivated in medium supplemented withmonosaccharides rather than with sucrose. The transcrip-tional activities of the CaMV 35S (pK5) or nos promoters(pLGVl 103) were unaffected by changes in the extracellularcarbohydrate source (data not shown, see also 35S core NPTin Table I). Repression of Sh promoter could be releasedby lowering the molarity of sucrose in the cultivation mediumfrom 175 mM to 57 mM which did not affect transcriptionalactivity of the CaMV 35S promoter (pK5 in Figure 2a).However, the hydrolysis of sucrose in the cultivation

medium cannot completely explain the activity profile of theSh promoter. As shown in Figure 2b, activity profiles similarto those obtained using sucrose medium were observed ifprotoplasts transfected with SP2014 +42 NPT were

cultivated on media supplemented with either glucose or

fructose (each 175 mM). A significant change in theconcentration of both extracellular monosaccharides was notmeasurable during the cultivation period (data not shown,see also accumulation of monosaccharides in Figure 2a).Therefore the delayed increase in Sh promoter activity isnot explained by changes in the extracellular carbohydrateconcentrations.

The increasing transcriptional activity is linked to cell

wall biosynthesisMicroscopic analysis indicated that during the 5 daycultivation period protoplasts started regeneration of cellwalls (see -DCB fluorescence photograph in Figure 3). Thesubstrate for the synthesis of most cell wall polysaccharidesincluding cellulose, is UDP-glucose, one of the productsresulting from the cleavage of sucrose by the enzyme sucrose

synthase, the product of the Sh gene. Therefore it was

tempting to test whether inhibition of cell wall formationaffects the activity of the Sh promoter as assayed by transientexpression experiments. After one large scale transfectionexperiment with SP20+42 NPT, protoplasts were sub-

+ULb

Fig. 3. Inhibition of cellulose biosynthesis blocks Sh promoter activity.Maize suspension protoplasts were cotransfected with SP20+42 NPTand pRTlOICAT, divided in two aliquots and cultivated on sucrosemedium (175 mM) for 5 days in the presence or absence ofdichlorobenzonitrile (DCB), a specific inhibitor of cellulosebiosynthesis. The delayed increase in NPTII activity of the Shpromoter is blocked by the herbicide. The activity of the cotransfected35S CAT gene (pRTIOICAT) indicates viability of the protoplasts inthe presence of DCB up to 5 days after transfection (CAT:chloramphenicol transacetylase enzyme, Cm: chloramphenicol, lCm,3Cm acetylated reaction products). The difference in cell wallformation dependent on the presence (+DCB) or absence (-DCB) ofthe herbicide is demonstrated by fluorescence staining of protoplasts 5

days after transfection (DAT). The entire surface is fluorescing inabsence of DCB, while only fluorescing spots are detectable inpresence of the cellulose biosynthesis inhibitor at higher exposuretimes. The chimeric SP2o+42 NPT gene has been used in thisexperiment because it was performed subsequently to the deletionalanalysis of the sucrose synthase promoter.

divided into two aliquots and cultivated for 5 days either innormal sucrose-containing media or in media supplementedwith 2,6-dichlorobenzonitrile (DCB), a specific inhibitor ofcellulose biosynthesis (Meyer and Herth, 1978). As shownin Figure 3 cell wall regeneration (+DCB fluorescencephotograph in Figure 3) was inhibited by the herbicide andthe late increase in Sh promoter activity was blocked. Thecontrol protoplasts cultivated in the absence of DCB showednormal cell wall regeneration (-DCB photograph inFigure 3) and increasing NPTII activity. The transientexpression of the cotransfected pRTIOICAT plasmiddemonstrates that the transcription of the CaMV 35Spromoter was not affected by the DCB treatment and thatprotoplasts were viable over the 5 day cultivation period inthe presence of the cellulose biosynthesis inhibitor.

Sequences near the transcription start are sufficientto mediate sugar type dependent and delayedpromoter activityAn unexpected result was obtained during analysis of 5'terminal promoter deletions. The longest deletion (SP20+42NPT), removing the TATA-box of the Sh promoter, was stillinhibited by high sucrose levels and its transcription increased

3449

3Cm1 Cm

Cm

1.1-1.7 1.0 -1.3

C.Maas, S.Schaal and W.Werr

between 2 and 5 days after transfection (see Table I andFigure 4a). For this deletion, the maize sequences fused infront of the NPTII marker gene consist of only 20 bp ofthe Sh upstream region and 42 bp of untranslated exon 1.The final NPTII activity observed in the transient expressionexperiments performed with this promoter deletion reachedsimilar levels to those obtained with the extended upstreamregions, e.g. SP2014+42 NPT. The comparison ofSP20+42 NPT to the promoterless NPTII cassette(pWWl 14-5 in Figure 4a) demonstrates that transcriptionalinitiation has to be mediated by the 62 bp Sh sequenceelement and not by sequences on the bacterial vector.

r .SP20+42 NPT

u 1 2 3 4 5 DAT

.... * . .

III

pWW 11 4 -5

OL C)c~~~~cE

n 0

OJQ 2 5 2 5FE 1 12

* X

*e *,.*

SP20 42 NPT

DAT

3Cm1Cm

C rrm

G A

m. .._so__..

_.

s;_1_

w-i4

w-

4._.O _

,.s_1F.*

235 229 n

277

~~~: _=i-291 4 2

291 SP f.t o c i '-

Fig. 4. (a) Time dependent NPTII activity profile of SP20+42 NPT incomparison with the promoterless pWW114-5 NPTII cassette. (b)RNA protection experiment performed with total cellular RNA isolatedfrom protoplasts transfected with the SP20+42 NPT plasmid (+) or

untransfected control protoplasts (-). The structure and the directionof the SP6 antisense transcript is shown below the autoradiograph, thelength of the protected bands has been determined according to a DNAsequence (G, A lanes shown) run in parallel. The bracket on themargin indicates the initiation sites between position +7 and +12relative to the main transcription start site used in the endosperm. Theband marked (*) extends to the 3' end of the SP6 transcript and canbe explained by unterminated readthrough transcripts or SP20+42NPT plasmid copies in the RNA preparation. (c) The exon 1

sequences in the Sh promoter NPTII fusions do not mediate a post-transcriptional response. The exon 1 +4 to +42 sequences werecloned into the transcription unit of pRTIOICAT as indicated inFigure la. In the upper autoradiograph the CAT activities ofpRTIOlCAT and pRTIO(+4+42)CAT are compared. CAT activitydecreased whether or not the maize leader sequences were present inthe transcription unit. Below the corresponding CAT lanes, expressionof cotransfected SP20+42 NPT shows the delayed increase in activity.Note that the insertion of the +4 to +42 sequences in the untranslatedleader sequence positively affected expression of the CAT gene. Cm:chloramphenicol and acetylation products (lCm, 3Cm), cont: proteinextracts from protoplasts treated in absence of DNA.

3450

This has been confirmed by RNase protection experiments(Figure 4b) which indicate transcripts initiating betweenposition +7 and + 12 relative to the main start site (+ 1)used in the endosperm (Werr et al., 1985). The prominentlonger band in the protection experiment coincides with thepoint of divergence between the SP6 transcript and thechimeric SP20+42 NPT construction. It may be explained,by some residual plasmid copies in the total RNApreparation, transcripts initiating in the pUC vectorsequences or unterminated readthrough transcripts. As thevery low background activity observed from the promoterlesspWWl 14-5 NPTII cassette argues against a significanttranslation of the NPTII coding region from such invadingtranscripts, we have not further analysed the nature of thisband.A post-transcriptional control mechanism mediated by the

Sh 42 bp untranslated leader sequence of exon 1 present inthe Sh promoter NPTII fusions seems unlikely for severalreasons. Firstly, the insertion of the exon 1 sequences (+4to +42) at a similar position in front of the CAT codingregion of pRTIOICAT did not change the time profile ofCAT expression (Figure 4c). CAT activity was high 2 daysafter transfection and decreased by day 5 while expressionof the NPTII enzyme from a cotransfected SP20+42 NPTconstruct increased between day 2 and day 5. Secondly, mostexon sequences are deleted from the chimeric SP20+6 NPTgene. The NPTII activity observed with SP20 +6 NPT wasconsiderably lower than obtained with SP20+42 NPT,however, transcription was still inhibited by sucrose andincreased late after transfection (Table I).A third argument against a control mediated by the Sh exon

1 is implied by the following promoter fusion experiments.In the SP2014-47core NPT chimeric gene the Shrunkentranscription start including the exon sequences is replacedby the CaMV 35S core promoter (SP2014-47core NPT seeFigure la). Transient expression of SP2014-47core NPT was6- to 7.5-fold higher in the presence of glucose than in thepresence of sucrose and increased up to 30-fold when NPTIIactivity was analysed after 5 days of cultivation. Expressionof the core promoter (35S core NPT in Figure la) wasunaffected by the carbohydrate source and cultivation period.From the promoter fusion experiments the presence of atleast one additional control element of similar function tothat near the transcription start site has to be postulatedupstream of -47 in the Sh promoter. The result alsoindicates that this type of transcriptional control can betransferred to other normally unaffected plant/viralpromoters.

DiscussionIs the sucrose synthase 1 promoter feedbackcontrolled?The activity of the Sh promoter in maize protoplasts asmonitored by transient gene expression was repressed byhigh extracellular sucrose concentrations but not bycomparable concentrations of glucose or fructose in theprotoplast medium. In addition, Sh promoter activityincreased sharply when the transfected protoplasts startedresynthesis of cell walls after 2-3 days of cultivation.Through the analysis of promoter deletions, a 26 bp region(SP 20+6) has been identified which responds negativelyto high sucrose concentrations and positively to cell wall

Feedback control element of maize

regeneration. Within this 26 bp sequence covering thetranscription start site of the maize Sh gene either twosequence elements, one responding to sucrose and one tocell wall synthesis, must be closely linked or both responses

are mediated via the same signal transduction chain.Such a chain could start with UDP-glucose, a reaction

product from the cleavage of sucrose by the enzyme sucrose

synthase, encoded by the Shrunken gene. Perhaps the Shpromoter is repressed at high intracellular UDP-glucoseconcentrations and active at low UDP-glucose levels. If DCBinhibits the transfer of glycosyl residues from UDP-glucoseto the growing cellulose chains UDP-glucose accumulatesand the activity of the Sh promoter stays low. The UDP-glucose level would also be affected at high extracellularsucrose concentrations if the disaccharide enters theprotoplasts and is cleaved intracellularily by the sucrose

synthase enzyme. Transcription and translation of theendogenous Sh gene has been verified in Northern andWestern blot experiments (data not shown). Therefore bothresponses could be mediated via the UDP-glucose level andthe signal involved could act at a single regulatory element.

If transcription of the Sh promoter is dependent on lowintracellular UDP-glucose levels the maize endosperm cellswhere the Sh gene is highly transcribed (Heinlein andStarlinger, 1989; Chen and Chourey, 1989) may be likeprotoplasts regenerating cell walls. Except, that in theendosperm cells UDP-glucose is a possible substrate for thesynthesis of starch (Tsai, 1974) rather than the biosynthesisof cellulose. The low UDP-glucose level at the high ratesof starch synthesis in the endosperm cells therefore wouldinduce Sh promoter activity. The translation of the Sh mRNAwould consequently result in an increased level of the sucrose

synthase enzyme supplying more UDP-glucose by thecleavage of sucrose.A similar feedback transcriptional control has been

described for the yeast SUC2 gene which encodes a secretedform of the invertase enzyme catalysing sucrose hydrolysis.Transcription of the SUC2 gene is repressed by glucose, a

product of enzymatic sucrose hydrolysis (Sarokin andCarlson, 1984). The experimental data in yeast providesevidence that regulation of the SUC2 gene involves signaltransduction with cyclic AMP as second messenger. CyclicAMP was initially shown to mediate Escherichia colicatabolite repression (review Magasanik and Neidhardt,1987) and subsequently to be involved in signal transmissionof a variety of eukaryotic genes (review: Roesler et al.,1988). Flanking the transcription start of the Sh gene are

six repetitive TCCC motifs (see Figure 5) exhibitingsequence homologies to AP2 binding sites mediating thecAMP response in the rat prolactin (CCCCTCCC, Cookeand Baxter, 1982) or tyrosine aminotransferase gene(TCCCTCCC, Becker et al., 1987). However, until now

cAMP has not been shown to be functional at the transcrip-tional level in plants. These fascinating problems of signaltransduction mechanisms in plants warrant further investi-gations.

The Sh transcription start is an autonomous controlelementApart from temporal regulation, the SP20+42 Sh elementinitiates transcription efficiently in the absence of furtherupstream sequences. The lymphocyte specific initiatorelement at the transcription start of the terminal deoxynucleo-

tidyltransferase gene is also sufficient for transcriptionalinitiation and control (Smale and Baltimore, 1989), whileadditional upstream elements are required for high transcrip-tional activity of the initiator. In contrast, high activity ofthe Sh transcription start does not appear to be dependenton upstream elements since sequences extending up toposition -2014 did not significantly alter or increase theexpression of the NPITH marker gene. This may be due, inpart, to the redundancy of regulatory elements, as indicatedby promoter fusion experiments (see SP201447 NPT inTable I). Sh promoter specific regulation was conferred on

the unresponsive CaMV 35S core promoter by fusion withthe SP2014-47 sequences. Therefore at least one additionalcontrol element exhibiting a similar function to that of thetranscription start region must be localized upstream ofposition -47. As other small Sh promoter elements or

individual binding sites for nuclear proteins (Springer et al.,1990) enhance transcriptional activity if inserted in front ofthe CaMV core promoter (C.Maas and W.Werr, unpub-lished), the type of regulation exhibited near the transcriptionstart of the Sh promoter seems to be dominant over thebinding of positively acting transcription factors.We have not directly compared the accuracy of transcrip-

tion initiation from sites used in SP20+42 NPT to the CAPsites used when additional upstream sequences are present,because our main interest is to identify the target site forsignal transduction. However, internal initiation of theSP20+42 Sh promoter element was detectable in RNaseprotection experiments as expected (Figure 4b). The CAPsites are scattered between position +7 and + 12 incomparison with the main transcription start site used inmaize endosperm (Werr et al., 1985). The shifted cap sitesmay be explained by the absence of the natural TATA-box(position -32 to -24) which has been shown to be involvedin positioning of the initiation complex in animal promoters(for example: Schatt et al., 1990). A three nucleotidedisplacement of Sh CAP sites was also detected in maizeroots after anaerobic induction of the Sh gene (Ortiz et al.,1988) which indicates that initiation may vary in vivo.As a first attempt to dissect the functions inside of the

SP20+42 element, we removed most of the exon 1

sequences. Transient expression experiments performed withSP20+6 NPT demonstrated that the 26 bp Sh element reactsin a similar way to the SP20+42 element, negatively to highsucrose levels or positively to cell wall regeneration. Thetotal NPTII activity, however, was considerably reduced incomparison with SP20+42 NPT. A positive effect on geneexpression due to the presence of Sh exon 1 sequences was

also observed with pRT101(+4+42)CAT (see Figure 4c)in which the exon was inserted at a similar position in frontof the CAT coding region. Presently we cannot distinguishwhether the exon sequences act at the transcriptional or post-

-20 i11 11A

+42

OTCCCTCTCCC5TCCCROAGAAACCCTCCCTCCCTCCTCCATT66ACT5CTT6CTCCCTGTT

NWiologv to gnAP2 CCCCTCCC--rol*ctin own

binding-sites TCCCTCCC-twosine aninot-nstoer-se gone

Fig. 5. Sequence of the sucrose synthase 1 transcription start (-20 to+42). The position of the initiation sites in SP 20+42 (see Figure 4b)are indicated by arrows. Sequence homologies to AP2 recognitionsequences are marked.

3451

C.Maas, S.Schaal and W.Werr

transcriptional level. More important is the finding that theSh untranslated leader sequences in front of the CAT codingregion do not alter the transcription profile of the CaMV35S promoter in the maize protoplasts. Therefore we feelquite sure that the 26 bp region surrounding the Sh transcrip-tion start (SP20+6) contains the transcriptional controlelement. In future experiments saturation mutagenesis of this26 bp region may allow identification of the protein bindingsite and subsequently the signal transducing nuclear protein.This will hopefully be a first step to understanding one ofthe signal transducing chains controlling the transcription ofgenes involved in carbohydrate metabolism.

Materials and methodsConstruction of chimeric genesAll sucrose synthase promoter chimeric gene constructions were cloned inthe same orientation into the polylinker region of pUC9 or 19. Cloningof the SP2014+42 NPT and the vectorless pWW1 14-5 NPTII cassette hasbeen described in Werr and Lorz (1986). The small SP20+42 or SP20+6promoter fragments have been inserted into the SniaI site of pWWl 14-5.The 35S core NPT gene was constructed by replacing an EcoRI-SphIfragment from pWW1 14-5 by a fragment from the vector pPCV821-neokindly provided by C.Koncz (Max-Planck-Institut fur Zuchtungsforschung,Koln). Fusion of the SP 2014-47 upstream sequences in front of the corepromoter was achieved by using the compatibility of a StyI restriction site(position -47) in the maize sequences and a XbaI site in front of the CaMVcore promoter. The plasmid pRT11(+4+42)CAT was constructed byinserting a synthetic oligonucleotide into the SmaI restriction site in theuntranslated leader sequence of pRTlOICAT (Prols et al., 1988).

Transient expression analysisProtoplasts were isolated from a maize suspension cell line, transfected witha 0.1 M MgCl2, 0.45 M mannitol, 20 mM HEPES, 25% PEG solutionand cultivated according to Maas and Werr (1989). To normalize individualtransfection experiments 50 ,g sucrose synthase promoter plasmids weremixed with 50 4g pRTlO1CAT and 100 Mg of sonicated calf-thymus DNAand incubated with 2 x 106 protoplasts in 300 ,ul. Each transfectionexperiment was carried out in triplicate, protoplasts were incubated in Petridishes at 28°C in the medium and for the time indicated in the figures.The three corresponding dishes were combined, protoplasts pelleted bycentrifugation, divided into two aliquots and cells lysed by sonication priorto CAT and NPTII activity assays.The estimation of CAT activity and protein concentration followed the

protocol described in Maas and Werr (1989), both values were used tonormalize differences in transfection efficiency. NPTII activity was assayedaccording to Reiss et al. (1984), protease activity was blocked by the additionof BSA (50 Mg) and PMSF (1 mM) to the extracts. Autoradiographs wereeither scanned densitometrically or the radioactive spots were cut out andanalysed in a scintillation counter.

RNA preparation and RNase protection analysisTotal RNA was prepared from protoplasts essentially as described byLogemann et al. (1987). For preparation of the antisense SP6 transcriptan EcoRI-PvuII fragment from the SP20+42 NPT chimeric gene wascloned into the pSP 64 vector (SmaI-EcoI, Melton et al., 1984). Thisfragment contains the 20 bp upstream region plus pUC polylinker sequencesand the 5' end of the bacterial NPTII coding sequence. SP6 transcripts wereprepared by using a SP6 transcription kit (Amersham, Code RPN 2006).Prior to the preparation of total RNA, protoplasts were incubated with DNaseI according to Werr and Lorz (1986) to remove extracellular plasmid copies.After RNA isolation 20Mg of total RNA was treated with DNase I (RNasefree; Boehringer Mannheim No. 776785) and5precipitated in the presenceof [32P]UTP-labelled RNA probe (1-2 x 10 c.p.m.). RNase protectionanalysis was performed essentially as described by Zinn et al. (1983).Hybridisation was performed at 45°C. Protected RNA fragments wereanalysed on denaturing polyacrylamide gels.

Carbohydrate determination and analysis of cell wallregenerationExtracellular carbohydrate concentrations were determined by test kitsavailable from Boehringer Mannheim GmbH (No. 139106). Cell wallregeneration was confirmed by fluorescence analysis with Tinopal CBS-X[CBS: disodium 4,4'-bis(2-sulfostyryl)biphenyl; Ciba-Geigy Corp.] asdescribed by Huang et al. (1986). Cellulose biosynthesis was inhibited by

3452

adding the herbicide DCB (2,6-dichlorobenzonitrile, Aldrich Chem.,D5,755-8) to the media N6-B 500 at 2 mg/l according to Meyer and Herth(1978).

AcknowledaementsSynthesis of oligonucleotides was kindly performed by Dr Bruno Gronenbom(Max-Planck-Institut fur Zuchtungsforschung, Koln). We thank Drs SarahGrant, Alfons Gierl, Andrew Davidson and Peter Starlinger for criticalreading of the manuscript. This work was supported by the DeutscheForschungsgemeinschaft (DFG) through SFB 274.

ReferencesAkazawa,T. and Okamoto,K. (1980) The Biochemistry ofPlants. Academic

Press Inc. Vol. 3. 199-220.Becker,P.B., Ruppert,S. and Schutz,G. (1987) Cell, 51, 435-443.Chen,Y.-C. and Chourey,P.S. (1989) Theor. Appl. Genet., 78, 553-559.Cooke,N.E. and Baxter,J.D. (1982) Nature, 297, 603-606.Chourey,P.S. and Nelson,O.E. (1976) Biochem. Genet., 14, 1041-1055.Creech,R.G. (1965) Genetics, 52, 1175-1186.Doehlert,D.C., Kuo,T.M. and Felker,F.C. (1988) Plant Physiol., 86,

1013-1019.Geiser,M., Weck,E., Doring,H.P., Werr,W., Courage-Tebbe,U.,

Tillmann,E. and Starlinger,P. (1982) EMBO J., 1, 1455-1460.Giaquinta,R.T. (1983) Annu. Rev. Plant Physiol., 34, 347-387.Griffith,S.M., Jones,R.J. and Brenner,M.L. (1987) Plant Physiol., 84,472-475.

Hain,R., Stabel,P., Czernilofsky,A.P., Steinbi/3,H.H., Herrera-Estrella,L.and Schell,J. (1985) Mol. Gen. Genet., 199, 161-168.

Heinlein,M. and Starlinger,P. (1989) Mol. Gen. Genet., 215, 441-446.Huang,C.-N., Cornejo,M.J., Bush,D.S. and Jones,R.L. (1986)

Protoplasma, 135, 80-87.Klosgen,R.B., Gierl,A., Schwarz-Sommer,Zs. and Saedler,H. (1986) Mol.

Gen. Genet., 203, 237-244.Kraus,K. (1988) The action of antisense RNA on the gene expression of

plant viruses. Ph.D. thesis, University of Cologne.Logemann,J., Schell,J. and Willmitzer,L. (1987) Anal. Biochem., 163,

16-20.Maas,C. and Werr,W. (1989) Plant Cell Rep., 8, 148-151.Magasanik,B. and Neidhardt,F. (1987) In Neidhardt,F.C. (ed.), Escherichia

coli and Salmonella typhimurium, Cellular and Molecular Biology.American Society for Microbiology, Washington, DC, 1318-1325.

McCarty,D.R., Shaw,J.R. and Hannah,L.C. (1986) Proc. Natl. Acad. Sci.USA, 83, 9099-9103.

Melton,D.A., Krieg,P.A., Rebagliati,M.R., Maniatis,T., Zinn,K. andGreen,M.R. (1984) Nucleic Acids Res., 12, 7035 -7056.

Meyer,Y. and Herth,W. (1978) Planta, 142, 253-262.Ortiz,D.F., Rowland,L.J., Gregerson,R.G. and Streommer,J.N. (1988) Mol.

Gen. Genet., 214, 135-141.Pr6ls,M., Topfer,M., Schell,J. and Steinbi(3,H.-H. (1988) Plant Cell Rep.,

7, 221-224.Reiss,B., Sprengel,R., Will,H. and Schaller,H. (1984) Gene, 3, 217-223.Roesler,W.J., Vandenbark,G.R. and Hanson,R.W. (1988) J. Biol. Chem.,

263, 9063-9066.Rowland,L.J., Chen,Y.C. and Chourey,P.S. (1989) Mol. Gen. Genet., 218,33-40.

Sarokin,L. and Carlson,M. (1984) Mol. Cell. Biol., 4, 2750-2757.Smale,S.T. and Baltimore,D. (1989) Cell, 57, 103-113.Springer,B., Werr,W., Starlinger,P., Bennert,D.C., Ziokolica,M. and

Freeling,M. (1986) Mol. Gen. Genet., 205, 461-468.Springer,B., Bellmann,R. and Werr,W. (1990) Mol. Gen. Genet., 221,94-101.

Schatt,M.D., Rusconi,S. and Schaffner,W. (1990) EMBO J., 9,481-487.Stitt,M. and Quick,P. (1989) Physiol. Plant., 77, 633-641.Tsai,C.Y. (1974) Biochem. Genet., 11, 83-96.Tsai,C.Y., Salamini,F. and Nelson,O.E. (1970) Plant Physiol., 46,299-306.

Werr,W. and Lorz,H. (1986) Mol. Gen. Genet., 202, 471-475.Werr,W., Frommer,W.-B., Maas,C. and Starlinger,P. (1985) EMBO J.,

4, 1373-1380.Zinn,K., DiMaio,D. and Maniatis,T. (1983) Cell, 34, 865-877.

Received on May 7, 1989; revised on August 2, 1990